Only use data from season 2013-2014

Number of games

The amount of Games
Var1 Freq
Playoffs 91
Regular Season 1221

Distribution of goals in different period time

Questions:

  1. penalty in games?

  2. the manpower of the begining of games in period 2 and 3 is 0?

## 
##      lose tied_lose  tied_win       win 
##       436       166       157       553

===== Count Matrix - Ratio =====

Group by Home Team Result(win/lose/tied win/tied lose) & GoalDiff

Time interval = 120 seconds

Win

Hometeam wins

For GoalDiff = 0 plot, the value of count decreases from perid 1 to 3. In period 1, the trend goes down more significantly then in period 2 and 3, the value of count is below average after period 1.

Hometeam lose

Hometeam tied_win

Hometeam tied_lose

Hometeam win(Regular season)

Hometeam lose(Regular season)

Hometeam tied_win(Regular season)

Hometeam tied_lose(Regular season)

Hometeam win(Playoffs)

Hometeam lose(Playoffs)

Hometeam tied_win(Playoffs)

Hometeam tied_lose(Playoffs)

Time interval = 60 seconds

Time interval = 30 seconds

Group by Home Team Result(win/lose/tied win/tied lose) & ManpowerDiff

Time interval = 120 seconds

Time interval = 60 seconds

Time interval = 30 seconds

Group by Home Team Result(win/lose/tied win/tied lose) & ManpowerDiff & GoalDiff

===== Count Matrix - After =====

Group by Home Team Result(win/lose/tied win/tied lose) & GoalDiff

Time interval = 120 seconds

Time interval = 60 seconds

Time interval = 30 seconds

Group by Home Team Result(win/lose/tied win/tied lose) & ManpowerDiff

Time interval = 120 seconds

Time interval = 60 seconds

Time interval = 30 seconds

===== Count Matrix - Before =====

Group by Home Team Result(win/lose/tied win/tied lose) & GoalDiff

Time interval = 120 seconds

Reward matrix

## # A tibble: 6 x 4
##   interval ManpowerDiff GoalDiff      p
##   <fct>           <dbl>    <dbl>  <dbl>
## 1 (0,120]            -1       -2 -0.650
## 2 (0,120]             0       -2 -0.694
## 3 (0,120]             1       -2 -2    
## 4 (0,120]            -1       -1 -0.658
## 5 (0,120]             0       -1 -0.379
## 6 (0,120]             1       -1  1.13

==============================

Analysis

  1. Different intervals - The trend of lines are almost same for different intervals, but smaller intervals give more details about the ups-and-downs of the count value.

  2. Different counting method - It doesn’t have very significant difference of fluctuation between ratio method and after method, the main trend for each plot is quite similar.

  3. Regular Season vs Playoffs - The plots of count values for Regular Season games are basically looks same as the plots for Regular Season + Playoffs games. For Playoffs games, although the data for Playoffs games are not that enough, but still can see the roughly trends. It approximatly flucduate similar as Regular Season.

home field advantage